LC0cA2 ^

Learning Environments: Now, Then, and Someday

multiple polyspiral design

Nine-fold Multiple-polyspiral @ 177 degrees

Complaints About Computers in Education

Microcomputers inexpensive enough for widespread use in education
became available in the late 70s in the USA — even some with enough capacity
to support conversational programming languages. Such micros seemed a
significant change from existing time-shared systems. Many educators saw the
promise of micros, and parents, in their eagerness to do what they imagined
best for their children, with personal contributions and political pressure led
school systems to invest heavily in that technology. Have those computers
produced results which justify the hope placed in them? I think not. The
disappointment expresses itself in four common complaints:

About computers in education, generally

THEME SPECIFIC COMPLAINT

Effect: Computers have not improved education.

Value: Computer experiences are inferior to real ones.

About learning environments:

Clarity: The notion is not clear and distinct.

Design: Nobody knows how to make them.

The specific complaints we will pursue here are those focussed on the themes of
effect, clarity and design. I conclude with some suggestions for future work.[1]

THE LIMITED IMPACT OF COMPUTERS IN EDUCATION

The introduction of microcomputers into the education system has
disappointed many people who had hoped their presence would engender reforms
leading to education both more congenial to children and more effective than
the norm of past generations. There has been no widespread recognition of any
such dramatic impact. Why? In a review of Computer Experience and Cognitive
Development
, Erik DeCorte noted:

“I point to the immediate connection between the book and the current inquiries
about using computers with children. Some have claimed that computer
experience, and the ability to program in particular, would influence in a
positive way the learning and thinking capacities of children. In contradiction
to the image produced by the rest of the available research literature (See
DeCorte & Verschaffel, 1985), Lawler’s study produces positive results
concerning the cognitive-effects hypothesis….”

Such negative outcomes as others report, when the result of thoughtful
experiments which are executed with care, have the proper function of
constraining the enthusiastic claims of the overly optimistic. On the other
hand, I am convinced that one reason for the difference of outcomes noted by
DeCorte is a consequence of different levels of detail of the studies. Too much
evaluative research has the flavor of I/O models: some INPUT should
produce some OUTPUT; some INSTRUCTION should produce some
OUTCOME. If the true orderliness of human behavior becomes evident only
when one looks very carefully at extremely tiny details, most experimental
efforts to assess computing’s impact will show negative results, unless they
examine the PROCESS between input and output, the PERSON between
instruction and outcome. Studying the knowledge and functioning of one mind in
detail permits a depth of understanding of the student’s mind and development
normally beyond the reach of research with a broader focus. We need to go
beyond evaluative studies of broad claims in order to advance our
understanding of human cognition, specifically in respect of the issues of the
malleability of the natural mind and of the long-term effects of specific
experiences on the lives of individuals: both, for me, are central issues for
the science of education.

Comments such as the preceding, though true and valuable, evade rather than
answer the question raised by DeCorte’s observation. It is the case that early
Logo claims looked to widespread results so obvious and striking that
corroborating or disambiguating experiments would not be required. No such
strong outcome has occurred. Computers, as introduced in schools, have not had
so beneficial an impact as their early proponents suggested they might. Let’s
reflect on this problem.

The Worst Case: The Problem is Not Solvable

Despite widespread research in several paradigms directed to improving
children’s mathematical competence through using computers, there is a general
impression, based on test results, that arithmetic skills have been
deteriorating over the past 25 years. One jocular suggestion for reacting to
this situation comes from “The Uses of Education to Enhance Technology” [2]

“It’s time to face the facts: all previous efforts at educational reform have
been failures. The harder we try, the more innovations we make, the dumber the
students get. This is eloquently pointed out by proponents of the ‘Back to
Basics’ movement in numerous riots and book burnings across the country.

The solution is clear. It is simply not possible to educate children. If
repeated attempts to improve the quality of education only make matters
worse, then the obvious way to make matters better is to try to degrade
the quality of education. In fact, carrying this argument to its logical
conclusion proves that the best educational reform would be to abolish efforts
at education altogether. This conclusion is hardly new and has been previously
argued by such thinkers as Holt and Illich. But they also foresaw the serious
impediments to this scheme. It would abolish the major value of the school
system, which is to supply employment and positions of power. . . . But now,
with modern technology, we see a way out of this dilemma. The solution is
absurdly simple: by placing computers in the schools, we can let the teachers
teach the computers and send the children home. . . . Specifically, we envision
an educational system in which each child is assigned a personal computer,
which goes to school in place of the child. (Incidentally, it should be noted
that the cost of such a personal computer is not large. Even at today’s prices
it is probably not much more than the average family would spend on a
catastrophic medical emergency)….”[3]

Those of us who are laughing through our tears can not escape the need for some
different way of dealing with the issue. One can try to take a broad view. It
is possible to believe that the problem has not been a “local” failure,
ascribable in some simple way to faulty research, slow technology transfer, or
intractable institutions.

An Explanation: Social Changes are Dominant

The problem may be profound and even could involve deterioration in the
learnability of common sense knowledge. Changes in the everyday world can
completely overwhelm our hopes to teach children skills we know they will need
later. Consider these observations (from Lawler, 1985) as an example of ways
in which social forces can radically alter the cognitive impact of domains of
common sense knowledge.

Vignette 55

“Since the beginning of the High School Studies Program, the children and I
have come to Logo to use the system from 8 to 10 a.m. The children have become
accustomed to mid-morning snacks. The favorite: apple pie and milk. At their
young age, Robby and Miriam get money from me, and we talk about how they
spend it. A piece of pie costs 59 cents. A half pint of milk is 32 cents. So
Miriam told me this morning, and these figures are familiar. As we got her
snack, I asked Miriam how much we would have to pay the cashier. After a few
miscalculations, she came to a sum of 91 cents and seemed confident it was
correct. I congratulated her on a correct sum and asked the cashier to ring up
our tab. “92 cents.”

“92 cents?” I asked the cashier to explain. She said the pie is 55 cents and
milk 30 cents, thus 85 cents and the tax, 7 cents. “See. Look at the table.”

I am at a complete loss as to how to explain this to Miriam. Not only is the 8
percent food tax dreadful in itself, but it is rendering incomprehensible a
primary domain of arithmetic that children regularly confront — paying small
amounts of money for junk food. Otherwheres, Miriam used “the tax” as a label
for the difference between what is a reasonable computation and what you
actually have to pay somebody to buy something….”

The observation suggests that a specific governmental policy has, as a side
effect, been making the world less sensible and harder to learn about. If, to
get accurate results, a child must learn to multiply and round (for computing a
percentage tax) before learning to add, he is in BIG trouble. If it does
no good to calculate correctly, because results will be adjusted by some
authoritatively asserted incomprehensible rule, why should one bother to be
over-committed to precision? If addition no longer adds up, what good is
arithmetic? If you can’t count on number, what can you can you count on? If
knowledge is not useful, why bother with it?

Complex technology may also be making the world less comprehensible, but the
effects are not uniform. Calculators and modern cash registers which compute
change obviate the need for much mental calculation. Contrariwise, it is
possible to argue that technology is making access to reading knowledge
easier.[4] These observations leave us with more
questions than answers, but the questions are addressable and significant ones:
to what extent is it possible to learn what one needs to know through everyday
experience? how do side effects of decisions by adults constrain or enhance
children’s ability to learn about the world in natural ways?

An Excuse: The Political Climate has been Adverse

After such observations, it is reasonable to ask how political decisions
— such as support for research — have influenced the use of computers in
education. For many years the federal government, through various agencies, was
a major supporter of research into technology for education. The impact of
the first Reagan budget — which proposed to reduce funding for research in
science education from $80M to $10M in one year — led to significant
demoralization of that community and deterioration of function within its
organizations [5]. The decimation of this
community was decidedly unhelpful and may have engendered some of the chaos
and superficiality of work evidenced as microcomputers were sold by the private
sector to the education community throughout the United States. On the other
hand, one must note that the more generous support provided by the French
Government’s founding of Le Centre Mondial pour l’Informatique et Ressource
Humaine
had no happier outcome, as noted by Paul Tate in Datamation.

“The Center intended to use microcomputers to take computing to the people
through educational workshops in both the developed and the developing world.
Field projects were set up in France and Senegal, and research schemes were
introduced covering interactive media, systems architecture, AI, user
interfaces, and medical applications. It was to be an international research
center independent of all commercial, political, and national interests.
Naturally, it failed. Nothing is that independent, especially an organization
backed by a socialist government and staffed by highly individualistic industry
visionaries from around the world. Besides, altruism has a credibility problem
in an industry that thrives on intense commercial competition. By the end of
the Center’s first year, Papert had quit, so had American experts Nicholas
Negroponte and Bob Lawler. It had become a battlefield, scarred by clashes of
management style, personality, and political conviction. It never really
recovered. The new French government has done the Center a favor in closing it
down. But somewhere in that mess was an admirable attempt to take high
technology, quickly and effectively, along the inevitable path into the hands
of the public
. The Center had hoped to do that in different countries. . . .
The Center is unlikely to be missed by many. Yet, for all its problems, it
made a brave attempt to prepare for some of the technical and market realities
of the next few years. We regret that such a noble venture met with such an
ignoble end.” [6]

An Excuse: Available Hardware has been Inadequate

There is no question that the introduction of computers in education
was a financial success — for some few companies — but the record with
respect to product engineering and the advancement of social goals is one of
nearly consistent failure. Consider, as an example, this brief review of the
development of Logo-capable microcomputers for education:

  • The GTI 3500 (a DEC LSI-11 with a Minsky-designed front end, the ‘2500’)
    was an interesting product that came to market too early. The 2500 implemented
    a special video-turtle primitive, spin (proposed by D. Hillis), which
    set the object rotating at a constant angular velocity. This machine would
    have been very useful to physics and engineering students had it survived.
  • Texas Instruments supported the development of Logo for the TI-99 at the MIT
    Logo Project. The turtle geometry component of the system was quite
    inadequate. The sprite graphics system, which originally had been an
    uninteresting feature of the hardware, proved in the end to be a liberating
    addition to the repertoire of tools which could be used for educating children
    with computers. The TI-99 captured a significant portion of the education
    market, and the company made good money with their product. Nonetheless, TI
    withdrew the product, even though their president wanted the corporation to
    remain active in that market. This decision had a radical, negative impact on
    the production of software for education. I have been told that many small
    educational software development efforts collapsed after this decision was
    taken.
  • The Apple II offered the best early versions of turtle geometry and list
    processing with Logo. Both it and the IBM PC were technically adequate systems
    for the time but were really more suited for use by junior high and older
    students. For use with younger children, the systems were, in fact, regressions
    from the graphics capabilities available in the TI-99.
  • Atari Logo offered two advances. With time-sliced “when demons”, the Atari
    permitted a technically primitive but intellectually deep form of
    multi-programming under user control. The four software sprites attempted to
    replace the expensive TI graphics hardware with simulated capabilities. The
    research lab was axed before their developmental projects came to fruition.
  • The Coleco Adam had the best mix of hardware and software function for
    education use, but the manufacture and assembly of the machine was beset with
    problems of quality control that were insurmountable.
  • The MSX microcomputers, some of which are the best Logo-capable systems
    commercially available today, are not imported into the United States because
    the major European producer Phillips believes (rightly, I suspect) that the
    education market has been saturated with Apple II’s and IBM PC’s.
  • The best generally available system for use with small children till now has
    been the Apple II with a plug-in sprite board containing the TI-99 graphics
    processor. The two main advantages provided by Apple Sprites were an increase
    in the memory size, permitting more complex collections of procedures to be
    assembled, and the addition of a drawing capability for the sprites. Recently,
    the board and software were discontinued, for both technical and marketing
    reasons.

This record will not convince anyone of the grand success of the private
sector in doing any more than making and losing money. With such volatile
markets, very little that takes thought or time can get done. Someday we will
have a stable computer product permitting the development of good educational
software for young children. That day is not yet. If better systems come
along, will they — like the MSX — fail either to reach the market or to sell
because the enthusiasm and capital of the public and the education systems have
been used up? It is quite possible that future opportunities have been
polluted through temerity and over-selling.

An Explanation: The Medium Has No Consumable Content

The disappointing impact of computers on education may be partly
explained by the lack of content addressable with the technology. Consider, in
contrast, the video cassette recorder. VCR’s have reached a “take-off” point
and now are present in over 30 percent of American homes. A key element in the
success of VCR technology — not only in the market but also in user
satisfaction with it — is the existence of a massive stock of material which
the VCR brought to a new level of accessibility. As a “follow-on” technology,
VCR’s reproduce for resale the production of 70 years of film and TV with
marginal conversion costs.

What existing material do micros have accessible? Ideas? Yes, but they must
be recoded for each new system unless microcoded emulation of predecessor
machines becomes common. More to the point, existing larger systems and
minicomputers typically have different purposes than did micros purchased for
education. To offset this limitation, conversational programming languages
suggested the possibility of extensive programming by end-users. What
precisely that meant and what has evolved from that hope is a theme discussed
in the various texts under the headings of Education and Computing.

Publication notes:

  • Written in 1986.
  • Published as Chapter 1 of Artificial Intelligence and Education, (Lawler and Yazdani, Eds.), Ablex, 1987.

Text notes:

  1. In Artificial Intelligence and Education (Lawler & Yazdani, 1987), DiSessa’s Chapter 3 addresses the value theme in his discussion of artificial worlds and real experience.
  2. This vicious and amusing document parodying the style of Papert’s early Logo proposals was written in 1977 within the MIT community by one “Admiral Turtle”, an author whose identity to this day remains in the fog.
  3. The preceding was written in 1977. If the Admiral were writing today, he might be tempted to substitute a more current evaluation, such as “The savings on student lunches alone would pay for the investment in two to three years.”
  4. See Chapter 5 (Lawler and Lawler) in Artificial Intelligence and Education, Ablex, (1987).
  5. For a revelation of the processes producing first, such draconian proposals, and then the budget deficits of Reaganomics, see D. Stockman’s personal history of the period, The Triumph of Politics.
  6. For a more extensive discussion of the controversies surrounding that center, see Dray and Menosky, 1983.


After Thoughts

a thinker something to think about

Technology is just stuff. What’s important is thinking about one’s experience. One ambition of the Logo project was to make stuff that young people would find worth thinking about. You see Rob on the left sitting like Rodin’s Thinker in the midst of Tinker Toys. What Has He Got To Think About ? The device on the right is a computer, made out of wooden Tinker Toys, by some of the marvelous young people who hung around the lab. Margaret Minsky, Danny Hillis, Gary Drescher, Ed Hardebeck, Brian Silverman, Steve Haines, Pat Soblavero, Leigh Klotz, and many others whose names escape me now. They were the real learning environment; with the founders and staff of the Logo lab, they all saw computing as a medium for expression and thought-full play.